A PoC AI Security Agent built to simulate and defend against prompt injection, identity spoofing, and role escalation in large language models.
- Prompt injection detection
- Role-based input filtering
- Azure Sentinel alerting integration
- Simulated attack scenarios
- Clone this repo
- Run
pip install -r requirements.txt - Launch with
python ai_agent.py
src/: AI agent & security logicdemo/: Attack scripts and walkthroughsdocs/: Architecture diagrams and flow
For educational and demonstration use only.